The Geography Of Generative AI

An economic advance that's targeting educated, high-skilled and well-paid workers, and what it means.

Source: gettyimages.com

A couple weeks ago, I purchased a new laptop. The laptop has Microsoft Copilot, the Microsoft AI tool launched in 2023. For kicks I thought I’d try it out. I asked Copilot to provide me with a draft on a topic for a future post. Now, I’ve tried ChatGPT in the past, so I had an idea of what to expect. However, I was blown away when Copilot immediately came back with a three-page, almost 1,000-word draft on the topic. The Copilot draft needed work; it was a factual statement rather than an essay, written in a voice that wasn’t targeted toward any particular reader or audience. Still, it formed the foundation for what I would later publish. It was definitely the kind of quick-serve research on a topic that can cut my writing time down significantly. Suddenly I saw the implications of artificial intelligence (AI) in a new light.

AI is moving so fast, it’s challenging everything we think we know about how automation affects work.

With each passing day, the global economy is moving closer to integrating AI into the economic fabric. Broadly, workers understand that it will impact work, just as earlier efforts at automation have impacted work in the past. We also understand that some places will be impacted more than others. Researchers at the Brookings Institution are putting some thought into this, and their findings so far are interesting.

Some recent research conducted by Mark Muro, Shriya Methkupally, and Molly Kinder of the Brookings Institution’s Brookings Metro research center is finding that AI will impact work more deeply and broadly than previously considered. Gains in generative AI are being made at a fantastic rate, with AI compiling and generating content that was once the purview of humans. Whereas earlier automation efforts targeted routine tasks usually handled by low-skilled, low-wage workers, generative AI is showing itself to be well-suited to take on work that relies on cognitive skills today. For example, as the Brookings report says, “think coders, writers, financial analysts, engineers, and lawyers.”

The quote above is immediately followed by this:

“And while generative AI puts at risk the “routine” tasks of customer service and clerical work (often handled by female-staffed call centers, customer service lines, and HR teams, for example), it is currently not equipped to handle the manual work of manufacturing, the skilled trades, construction, and many in-person service industries.”

In other words, generative AI is coming after knowledge workers.

This led to Brookings Metro taking another tack on AI’s economic impact – what exactly will be the geography of generative AI? What metro areas would have the greatest exposure to generative AI, meaning the possible displacement of workers?

The Brookings Metro researchers reviewed occupation-specific “exposure” data supplied by ChatGPT creator OpenAI. They found that AI exposure increases (positively or negatively) as wages increase. The team made it clear that “exposure” doesn’t necessarily mean “worker displacement”. They allow for the fact that some jobs will be “augmented” by generative AI, enhancing the productivity and capability of many workers. Still, there’s a strong correlation that as education levels and the need for cognitive skills rises, so does potential exposure to generative AI. This graph from the report illustrates the correlation:

Their next step was to analyze exposure impact at the metro area level and at the county level. By analyzing data on local share of employment by occupation, by metro area and county level, the Brookings team was able to determine which metro areas and counties might have greater exposure to AI. To no one’s surprise, the metro areas and counties with the largest proportion of well-paying office jobs filled with highly educated workers reliant on cognitive skills had the greatest exposure to generative AI. The graph below highlights this:

The greatest exposure to AI will come to the coastal cities that blossomed as the knowledge economy expanded over the last 30 years. Using the Brookings data and analysis, Silicon Valley (the San Jose-Sunnyvale-Santa Clara metro area on the upper right) is the outlier for AI exposure, in terms of number jobs exposed to it, and the average annual pay of those jobs. San Francisco, Seattle and New York City are also vulnerable to AI exposure.

Metros located on the bottom left of the graph, however, are far less vulnerable. Las Vegas, a metro heavily reliant on tourism and hospitality, shows up in that quadrant, as do Peoria, IL and Massillon, OH, a couple of manufacturing dominant Midwestern cities.

As with every innovation that sparks economic change, some metro areas and their economies will benefit from this, while others won’t. How will it impact metros? At this point I see it as something akin to manufacturers moving factories to Southern cities in the ‘80s and ‘90s to take advantage of lower taxes and lower levels of labor union organization. That created a dispersion of manufacturing jobs throughout the country, after decades of being concentrated in the nation’s midsection. Workers followed the factories (even in reduced numbers, because of automation), and strengthened Southern economies.

What goes around comes around.